Objective Polymorphism of the androgen receptor (AR) has been related to various pathophysiological conditions, such as osteoporosis and infertility. The objectives of this study were to evaluate the frequency of distribution in a normal Italian population and to assess CAG repeats (CAGr) in other conditions, such as hypoandrogenism, potentially influenced by AR polymorphism. Patients and measurements CAGr polymorphism was determined in a group of 91 healthy normoandrogenized subjects, 29 hypoandrogenized patients (hypoplasia of prostate and seminal vesicles, reduced beard or body hair, etc.) and 29 infertile patients by direct sequencing. Results The mean (+/- SD) number of CAG repeats [(CAGr)n] was 21.5 (+/- 1.7) in the control group, 21.4 (+/- 2.0) in the infertile patients and 24.0 (+/- 2.9) in the hypoandrogenic males. The difference was statistically significant between this last group and the other two (P < 0.0001), while there was no difference between normal controls and infertile patients. The frequency distribution showed a shift towards higher CAG length in hypoandrogenized patients compared to controls and infertile patients. If we used a cut-off point of 24.9 (2 SD above the mean), the percentage of patients with 25 or more CAGr repeats was 38% among hypoandrogenized patients, 7% among infertile patients and 5% among the control group. In hypoandrogenized subjects (CAGr)n correlated slightly with testis and prostate volume. The number of CAG repeats was not associated with any of the hormonal parameters, including testosterone, evaluated in the three groups. Conclusions Our normal population, representing subjects from Central Italy, is superimposable on other European populations with regard to (CAGr)n distribution. Hypoandrogenic males have a shift in the frequency distribution towards longer (CAGr)n. Infertile patients are not statistically different from the control group. These findings suggest that, given the same amount of circulating testosterone, as in our hypoandrogenized and control group, the final net androgenic phenotypical effect is due to AR polymorphism.